- Email: [email protected]
Transamniotic stem cell therapy (TRASCET) in a leporine model of gastroschisis
Journal of Pediatric Surgery xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg
Transamniotic stem cell therapy (TRASCET) in a leporine model of gastroschisis☆ Christina Feng, Christopher D Graham, Hester F. Shieh, Joseph A Brazzo III, John Patrick Connors, Lucas Rohrer, Alexander Papadakis, David Zurakowski, Dario O Fauza ⁎ Department of Surgery, Boston Children's Hospital and Harvard Medical School, Boston, MA
a r t i c l e
i n f o
Article history: Received 27 September 2016 Accepted 20 October 2016 Available online xxxx Key words: Amniotic mesenchymal stem cells Transamniotic stem cell therapy Gastroschisis Fetal stem cells
a b s t r a c t Background/purpose: Transamniotic stem cell therapy (TRASCET) with amniotic fluid mesenchymal stem cells (afMSCs) has been shown to mitigate bowel damage in a rodent model of gastroschisis. As a prerequisite to clinical translation, we sought to study TRASCET in a larger animal model. Methods: New Zealand rabbit fetuses (n = 64) with surgically created gastroschisis were divided into three groups. One group (untreated) had no further manipulations. Two groups received volume-matched intraamniotic injections of either saline or a suspension of afMSCs. Nonmanipulated fetuses served as controls. Histomorphologic measurements of intestinal damage, along with biochemical profiling of inflammation markers, were performed at term. Statistical comparisons were by Fisher's exact test, ANOVA and the Wald test (P b 0.05). Results: Overall survival was 62.5%. Segmental and total intestinal wall thicknesses were significantly decreased in the afMSC group compared with the untreated and saline groups (all P b 0.001), with no significant differences between untreated and saline groups (P = 0.24 to 1.00, depending on layer). Muscularis and serosal layers were significantly thicker in the afMSC group than in normal controls (P = 0.045 and P b 0.001, respectively). Conclusions: Concentrated intraamniotic injection of afMSC lessens, yet does not prevent, intestinal damage in a leporine model of gastroschisis. TRASCET may become a valuable strategy in the management of gastroschisis. Level of evidence: N/A - animal/experimental studies. © 2016 Elsevier Inc. All rights reserved.
Transamniotic stem cell therapy (TRASCET) has recently emerged experimentally as a practical paradigm for the treatment of different congenital anomalies [1–3]. It is based on the amplification of a biological role played by naturally occurring amniotic fluid mesenchymal stem cells (afMSCs), which we have previously shown to contribute to normal fetal tissue repair [4]. The administration of concentrated suspensions of afMSCs into the amniotic cavity, as per the TRASCET principle, has been shown to decrease intestinal damage in a rodent model of gastroschisis [3]. The eventual clinical translation of this therapeutic concept necessarily demands testing its efficacy in larger animal models. In this study, we sought to study TRASCET in a leporine model of gastroschisis.
☆ Authors' contributions:Data acquisition: CF, CDG, HS, JAB, JPC, LR, AP, and DOF.Analysis and data interpretation: CF, CDG, HS, DZ, and DOF.Drafting of the manuscript: CF, DZ, and DOF.Critical revision: CF, DZ, and DOF. ⁎ Corresponding author at: Boston Children's Hospital, Department of Surgery, 300 Longwood Avenue - Fegan 3, Boston, MA 02115. Tel.: +1 617 919 2966; fax: +1 617 730 0910. E-mail address: [email protected] (D.O. Fauza).
1. Methods This study was approved by the Boston Children's Hospital Institutional Animal Care and Use Committee under protocol #14-11-2852. 1.1. Donor afMSC processing Donor cells consisted of previously banked afMSCs obtained from normal heterogeneic New Zealand does time-dated at 23–24 days of gestation (E23–24; term = 32–33 days). Cells were isolated and expanded based on previously described methods [5]. Fluorescenceactivated cell sorting (FACS) analysis was used to confirm a mesenchymal progenitor identity using the Vantage SE cell sorter (BD Biosciences, San Jose, CA) and unconjugated mouse monoclonal antibodies previously validated for use in rabbits, namely CD44 (Antigenix America, Inc., Huntington Station, NY), CD45 (Antigenix), and CD90 (Life Technologies/Thermo Fisher, Carlsbad, CA). A mouse isotype immunoglobulin control was used to exclude nonspecific staining. Phenotyped and expanded cells were labeled by fluorescent nanocrystal technology using the Qtracker® cell labeling kit 525 (Life Technologies, Chicago, IL) emitting a wavelength of 525 nm with excitation at 405–485 nm, as per the manufacturer's instructions [6]. Green fluorescence was observed in up to approximately 90% of cells
http://dx.doi.org/10.1016/j.jpedsurg.2016.10.016 0022-3468/© 2016 Elsevier Inc. All rights reserved.
Please cite this article as: Feng C, et al, Transamniotic stem cell therapy (TRASCET) in a leporine model of gastroschisis, J Pediatr Surg (2016), http://dx.doi.org/10.1016/j.jpedsurg.2016.10.016
2
C. Feng et al. / Journal of Pediatric Surgery xxx (2016) xxx–xxx
postincubation using an EVOS® FL Color Imaging System microscope fitted with an on-board computer and integrated imaging software (Life Technologies, Carlsbad, CA). 1.2. Gastroschisis creation and treatment Eighteen time-dated pregnant New Zealand white does were housed individually under standard dark/light cycling conditions and fed a normal diet ad libitum. On E23, they were anesthetized with 1.5% to 2.5% isoflurane (Abbot Laboratories, North Chicago, IL) in 100% oxygen and given 20 mg/kg of cefazolin intravenously. Using sterile technique, the bicornuate uterus was partially exposed through a midline laparotomy. Depending on the number of fetuses present, one to two fetuses on the most distal aspect of each uterine horn were manipulated (i.e. 2–4 fetuses per doe). Under 2.5× magnification, a 1-cm hysterotomy was made, previously bordered by four equally spaced 5-0 Prolene (Ethicon, Somerville, NJ) stay sutures incorporating the myometrium and gestational membranes. Both hind limbs were exteriorized, while the remainder of the fetus was kept within the amniotic sac. A 3-mm full thickness abdominal defect was then created at either the right or the left lower quadrant using micro surgery scissors. This was followed by gentle blunt probing of the incision, so as to dilate the abdominal opening and allow for evisceration of bowel loops. Once evisceration was confirmed, the fetus was completely returned to the amniotic cavity and the hysterotomy was closed with running 5-0 Prolene suture (Ethicon). The uterus was returned to the abdomen and the laparotomy was closed in two layers with 3-0 Vicryl (Ethicon) and 5-0 Monocryl (Ethicon) running sutures. Animals were allowed to recover and proceed to term. A total of 64 fetuses with surgically created gastroschisis were divided into three groups. One group underwent no further manipulations (n = 22). Two treatment groups received volume-matched (2–3 mL) intraamniotic injections of either plain phosphate-buffered saline (PBS, n = 22) or a suspension of the previously processed heterologous rabbit afMSCs, at 2 × 106 cells/mL in PBS (n = 20), after a comparable amount of amniotic fluid had been removed from the amniotic cavity. The intraamniotic infusions were performed just prior to completion of the closure of the hysterotomy using an 18-gauge angiocatheter (Becton Dickinson, Franklin Lakes, NJ). The saline group was required for two reasons. One is the fact that the afMSCs were delivered in a saline suspension, which by itself could have had an effect. The other is the previous demonstration by other groups that saline alone may minimize bowel damage in the rabbit and other animal models, which has served as basis for amnioexchange having been attempted clinically. An additional group of nonmanipulated fetuses was used as normal controls (n = 10). 1.3. Histological analysis All does were euthanized just prior to delivery on E32–33 using 110 mg/kg of Fatal-Plus™ solution (Vortech, Dearborn, MI). The laparotomy was reopened and the uterus exposed. The amniotic cavities of the manipulated fetuses were incised and the fetuses inspected for the presence of gastroschisis. When present, the eviscerated portion of the intestine was excised en bloc. Approximately half of the specimen was fixed in 4% formaldehyde, paraffin embedded, and processed for regular histology with hematoxylin–eosin (H&E) staining. The other half was processed for biochemical analyses (details below). For nonmanipulated fetuses (normal group), a midline laparotomy was made and the intestines were removed en bloc from the pylorus to the rectum. Similarly in these fetuses, a portion of the normal intestine was fixed for H&E staining and another portion was processed for biochemical analyses. All histological analyses were performed using an EVOS® XL Core Imaging System microscope fitted with an onboard computer and integrated imaging software (Life Technologies). Quantitative histomorphologic measurements previously validated as surrogates of bowel damage in experimental gastroschisis [7,8] were
performed by three examiners blinded to the groups. Segmental thickness measurements of intestinal mucosa, muscularis, and serosa layers were performed separately in pixels at 200 × magnification using Image J software (NIH, Bethesda, MD) on at least four cardinal sections of axial slides of multiple intestinal loops. Four to 13 bowel loops per fetus were individually analyzed. Screening for labeled cell engraftment was performed in unstained slides using an EVOS® FL Color Imaging System microscope fitted with an on-board computer and integrated imaging software (Life Technologies). 1.4. Biochemical analysis Additional intestinal tissue samples from all groups were frozen at −80 °C at the time of procurement for measurements of myeloperoxidase (MPO), malondialdehyde (MDA), and interferon-gamma (IFN-γ) activities – all markers of inflammation. For the MPO and MDA colorimetry assays, samples were homogenized using the Bullet Blender Storm 24 (Next Advance, Averill Park, NY) and measured for absorbance at 412 nm and 532 nm, respectively, using a FLUOstar Omega microplate reader (BMG Labtech, Cary, NC) as per the manufacturer's protocol (Sigma-Aldrich, St Louis, MO). For the IFN-γ assay, sample homogenates were additionally prepared with 4 volumes of Dulbecco's PBS without Ca 2 + and Mg 2 + (Sigma-Aldrich) and 0.1% Triton™ X-100 (Fisher BioReagents, Pittsburgh, PA). Absorbance was at 450 nm on the FLUOstar Omega microplate reader using the RayBio® Rabbit IFN-γ ELISA kit per the manufacturer's protocol (RayBiotech, Inc., Norcross, GA). 1.5. Statistical analysis Fetal survival rates were compared using Fisher's exact test for binary proportions. Multiple bowel loops from each of the four comparison groups (normal, untreated, saline, and afMSC) were used in nested analysis of variance (ANOVA) comparisons of segmental and total intestinal wall layer thicknesses. A generalized estimating equation (GEE) approach with repeated measures was used in order to account for correlated data, e.g. multiple thickness measurements within the same fetus sample. Statistical significance was determined using the Wald test. Thickness data conformed to a normal Gaussian-shaped distribution, and thus, thickness measurements were presented using mean and standard error of the mean (SEM). Two-tailed values of P b 0.05 were considered statistically significant. Data analysis was performed using IBM SPSS software (SPSS version 23.0, IBM, Armonk, NY). 2. Results Of the 64 fetuses that underwent gastroschisis creation, 40 were viable at euthanasia (overall survival of 62.5%), of which 36 had a gastroschisis defect (Fig. 1). There was a statistically significantly higher rate of survival in the saline group (n = 20/22, 90%) when compared with the untreated (n = 11/22, 50%; P = 0.007) and the afMSC (n = 9/20, 45%; P = 0.002) groups. Survival was not significantly different between the untreated and afMSC groups (P = 0.77). Histomorphologic comparisons were performed in a total of 256 bowel loops distributed as follows: 87 loops in the normal control group, 67 in the untreated group, 54 in the saline group, and 48 in the afMSC group. There were no significant differences between the groups in the median number of bowel loops analyzed per fetus (P = 0.23). Segmental intestinal thicknesses of serosal, muscularis, and mucosal layers, as well as total intestinal wall thickness, were all significantly decreased in the afMSC group compared to those of the untreated (P b 0.001) and saline (P b 0.001) groups (Figs. 2 and 3). Total intestinal wall thickness and mucosal thickness of the afMSC group were not significantly different from those of normal controls (P = 0.26 and P = 1.00, respectively); however, both the muscularis and the serosal layers were increased in the afMSC group when compared to normal controls
Please cite this article as: Feng C, et al, Transamniotic stem cell therapy (TRASCET) in a leporine model of gastroschisis, J Pediatr Surg (2016), http://dx.doi.org/10.1016/j.jpedsurg.2016.10.016
C. Feng et al. / Journal of Pediatric Surgery xxx (2016) xxx–xxx
3
Fig. 1. Representative gross views of rabbit fetuses with gastroschisis at euthanasia. (A) Untreated group; (B) saline-treated group; and (C) afMSC-treated group.
(P = 0.045 and P b 0.001, respectively). Comparisons between the saline and untreated groups showed no significant differences in mucosa (P = 0.24), muscularis (P = 0.33), serosa (P = 0.90), and total (P = 1.00) intestinal wall thickness. Both the untreated and the saline groups had significantly increased segmental (all layers) and
total intestinal wall thickness compared to normal controls (P b 0.001 for all comparisons). Nanocrystal-labeled afMSCs were visualized sparsely throughout the intestine in the afMSC group, albeit not in the luminal epithelium (Fig. 3).
Fig. 2. Histological segmental and total intestinal wall thickness measurements were significantly decreased in the afMSC group compared with the untreated and saline groups (all P b 0.001), with no significant differences between untreated and saline groups (P = 0.24 to 1.00, depending on layer). However, muscularis and serosal layers were significantly thicker in the afMSC group than in normal controls (P = 0.045 and P b 0.001, respectively).
Please cite this article as: Feng C, et al, Transamniotic stem cell therapy (TRASCET) in a leporine model of gastroschisis, J Pediatr Surg (2016), http://dx.doi.org/10.1016/j.jpedsurg.2016.10.016
4
C. Feng et al. / Journal of Pediatric Surgery xxx (2016) xxx–xxx
Fig. 3. Representative histological views of the exposed intestine in (A) untreated, (B) saline, and (C) afMSC groups. The total bowel wall, serosal, muscular, and mucosal thicknesses appeared different in the afMSC group when compared with the untreated and saline groups, as reflected in the quantitative measurements shown in Fig. 2. However, they were not completely normal, as illustrated in (D) normal fetal intestine. H&E, 200× magnification. (E and F) Nanocrystal-labeled afMSCs were visualized sparsely throughout the intestine, except for the mucosal epithelium, in the cell-treated group. Unstained, 100× and 200× magnifications.
There were no significant differences in MPO (P = 0.82), MDA (P = 0.86), and IFN-γ (P = 0.17) activities across all groups. Specifically, concentrations detected were as follows, described as median values with interquartile range. For MPO - normal controls: 0 (0–19) mU/mL; untreated: 9 (0–16) mU/mL; saline: 3 (0–17) mU/mL; and afMSC: 13 (0–17) mU/mL. For MDA - normal controls: 0.39 (035–0.56) nmol/uL; untreated: 0.4 (0.29–0.48) nmol/uL; saline: 0.34 (0.24–0.51) nmol/uL; and afMSC: 0.39 (0.22–0.54) nmol/uL. For IFN-γ - normal controls: 0.48 (0.44–0.53) ng/mL; untreated: 0.54 (0.49–0.58) ng/mL; saline: 0.5 (0.47–0.55) ng/mL; and afMSC: 0.56 (0.49–0.63) ng/mL. 3. Discussion The prevalence of gastroschisis has increased significantly over the past two decades, for reasons as yet unknown. The most recent largescale data show its prevalence at almost 5 per 10,000 live births [9]. While multiple advances in perinatal care have led to exceedingly low mortality from isolated gastroschisis, little has changed in the significantly high morbidity rates associated with this disease, long known to derive largely from the prolonged exposure of the eviscerated intestine to amniotic fluid [10–12]. To date, no prenatal intervention, including deliberate preterm delivery, has been clinically proven to lessen intestinal damage, improve bowel function, or decrease overall morbidity. Although several approaches to minimize intestinal damage have been reported experimentally, a cell-based strategy had yet to be described until our recent study on TRASCET in a rodent model [3]. The TRASCET concept is centered on a primarily quantitative amplification of a naturally occurring biological process, somewhat analogous to the principles governing different forms of blood transfusions. Certainly, clinical translation of such a concept could not be properly justified without it being tested and refined in larger animal models. Although the sheep model is arguably the ultimate preclinical proving ground for this application, a rapid transition from the rodent to the exceedingly costly and resource-intensive ovine model would be less sensible than a more progressive one using an intermediate species, such as rabbits. Rabbits are members of the taxonomic order Lagomorpha, widely considered as a large animal model. Importantly in this case, the native amniotic fluid volume of rabbits is several-fold larger than that of rodents. More specifically, the volume of intraamniotic injections required for TRASCET in rabbits is 30 to 60 times larger than
that required for rats, depending on the disease process studied. The fetal rabbit model we used was essentially based on the first one described by Dr. Jens Rosenkrantz and colleagues in the 1970s, with few modifications [13]. While the leporine model is cost effective and carries a reasonable fetal survival rate, the leporine gestation is too short to allow for the use of autologous afMSCs, hence our use of banked heterologous cells. We saw no evidence of inflammation or any untoward reaction derived from the use of heterologous cells, as it was the case in the similar heterologous rodent model. At the same time, the exposed bowel in afMSCtreated animals was not entirely normal. We cannot rule out the possibility that autologous cells could have an impact on such results. The reasons for the statistically significantly higher survival rate in the saline group are not entirely clear, yet they may have derived from the fact that most animals in that group were the last ones to be manipulated, therefore after the steeper portion of the surgical learning curve had passed. The lack of statistically significant differences in the comparisons of segmental intestinal wall layer thicknesses between the saline and untreated groups is in accordance with both our results in the rodent model and the clinical experience with amnioexchange. Patients submitted to amnioexchange in the third trimester have not shown improvements in the degree of intestinal damage, time to full enteral feeds, or length of hospitalization [14]. In the present study, afMSC-based TRASCET did not completely reverse the bowel damage associated with gastroschisis, yet it did minimize it. The underlying mechanisms behind this outcome remain to be determined. Mesenchymal stem cells have long been known for possessing immunomodulatory and anti-inflammatory activities, which may well be germane to the effects we observed [15–17]. Nonetheless, we did not see significant differences between the groups in the activity of the inflammatory markers measured. Although molecules and signals other than those measured may be at play, the absence of at least certain inflammatory changes is actually in accordance with other studies, which also did not show variations in selective markers of inflammation in the setting of gastroschisis [14,18]. Besides further investigation into other components of the inflammatory response, such as quantitative and qualitative analyses of leukocyte activity and of connective tissue proliferation and/or fibrin deposition, among others, scrutiny into the status of native intestinal stem cells should be
Please cite this article as: Feng C, et al, Transamniotic stem cell therapy (TRASCET) in a leporine model of gastroschisis, J Pediatr Surg (2016), http://dx.doi.org/10.1016/j.jpedsurg.2016.10.016
C. Feng et al. / Journal of Pediatric Surgery xxx (2016) xxx–xxx
included, given that administered stem cells, including MSCs, are known to often exert a chaperone effect upon the local stem cell population, independently of inflammatory mediation. At the same time, we had to prioritize previously validated methods of comparison of bowel damage in models of gastroschisis, which, to our knowledge, are based on the methods used in the present study, this being the rationale for the methodology used. It was not our goal, nor was it possible, to understand all the mechanisms behind the protection afforded by the afMSC delivery, which we do intend to continue to investigate. Furthermore, as much as the mechanisms of afMSC-derived protection of the exposed bowel have yet to be clarified, the donor cell homing routes in that setting also remain to be fully elucidated and may not necessarily involve only direct seeding from the amniotic cavity. Of note, in the rabbit model, donor cells had to be administered at the time of creation of the defect, therefore before any bowel damage had actually taken place. This may have explained the sparse afMSC engraftment pattern observed. The timing of this model also does not allow for repeat cell injections, which could be envisioned clinically as means to enhance therapeutic benefit. At the same time, it could also be argued that, despite the unavoidable short time afforded by this model, there was still an effect, even if not a full reversal of the bowel damage. We expect the rodent model of gastroschisis to continue to be instrumental to further developments and to be pursued in parallel to larger animal models. The latter are not conducive to mechanistic investigations because of the notorious unavailability of supplies for pathway-specific analyses. Importantly, this and the other previously reported animal models of this disease have not faithfully and consistently reproduced the unique component of prenatal narrowing of the fascial defect present in the human condition, which plays a significant role in the mechanism of bowel damage besides the chemical insult derived from the exposure to amniotic fluid. It remains unclear whether TRASCET could have any impact on that component of the pathophysiology of this disease. Notwithstanding the inherent limitations of the leporine model of gastroschisis, it constitutes a helpful next step toward the eventual clinical translation of TRASCET. The present results showing that concentrated intraamniotic injection of afMSCs lessens intestinal damage in that model add further support to the notion that TRASCET may become a valuable strategy in the management of gastroschisis.
5
References [1] Dionigi B, Ahmed A, Brazzo III J, et al. Partial or complete coverage of experimental spina bifida by simple intra-amniotic injection of concentrated amniotic mesenchymal stem cells. J Pediatr Surg 2015;50(1):69–73. [2] Dionigi B, Brazzo III JA, Ahmed A, et al. Trans-amniotic stem cell therapy (TRASCET) minimizes Chiari-II malformation in experimental spina bifida. J Pediatr Surg 2015; 50(6):1037–41. [3] Feng C, Graham CD, Connors JP, et al. Transamniotic stem cell therapy (TRASCET) mitigates bowel damage in a model of gastroschisis. J Pediatr Surg 2016;51(1): 56–61. [4] Klein JD, Turner CG, Steigman SA, et al. Amniotic mesenchymal stem cells enhance normal fetal wound healing. Stem Cells Dev 2011;20(6):969–76. [5] Klein JD, Fauza DO. Amniotic and placental mesenchymal stem cell isolation and culture. Methods Mol Biol 2011;698:75–88. [6] Muller-Borer BJ, Collins MC, Gunst PR, et al. Quantum dot labeling of mesenchymal stem cells. J Nanobiotechnol 2007;5:9. http://dx.doi.org/10.1186/1477-3155-5-9. [7] Bittencourt DG, Barreto MW, Franca WM, et al. Impact of corticosteroid on intestinal injury in a gastroschisis rat model: morphometric analysis. J Pediatr Surg 2006; 41(3):547–53. [8] Hakguder G, Olguner M, Gurel D, et al. Induction of fetal diuresis with intraamniotic furosemide injection reduces intestinal damage in a rat model of gastroschisis. Eur J Pediatr Surg 2011;21(3):183–7. [9] Jones AM, Isenburg J, Salemi JL, et al. Increasing prevalence of gastroschisis – 14 states, 1995-2012. MMWR Morb Mortal Wkly Rep 2016;65(2):23–6. [10] Langer JC, Bell JG, Castillo RO, et al. Etiology of intestinal damage in gastroschisis, II. Timing and reversibility of histological changes, mucosal function, and contractility. J Pediatr Surg 1990;25(11):1122–6. [11] Langer JC, Longaker MT, Crombleholme TM, et al. Etiology of intestinal damage in gastroschisis. I: effects of amniotic fluid exposure and bowel constriction in a fetal lamb model. J Pediatr Surg 1989;24(10):992–7. [12] Srinathan SK, Langer JC, Blennerhassett MG, et al. Etiology of intestinal damage in gastroschisis. III: morphometric analysis of the smooth muscle and submucosa. J Pediatr Surg 1995;30(3):379–83. [13] Sherman NJ, Asch MJ, Isaacs Jr H, et al. Experimental gastroschisis in the fetal rabbit. J Pediatr Surg 1973;8(2):165–9. [14] Midrio P, Stefanutti G, Mussap M, et al. Amnioexchange for fetuses with gastroschisis: is it effective? J Pediatr Surg 2007;42(5):777–82. [15] Barry F, Murphy M. Mesenchymal stem cells in joint disease and repair. Nat Rev Rheumatol 2013;9(10):584–94. [16] Phinney DG, Prockop DJ. Concise review: mesenchymal stem/multipotent stromal cells: the state of transdifferentiation and modes of tissue repair – current views. Stem Cells 2007;25(11):2896–902. [17] Wang Y, Chen X, Cao W, et al. Plasticity of mesenchymal stem cells in immunomodulation: pathological and therapeutic implications. Nat Immunol 2014;15(11):1009–16. [18] Goncalves FL, Bittencourt DG, Velloso LA, et al. Corticosteroid effect upon intestinal and hepatic interleukin profile in a gastroschisis rat model. Acta Cir Bras 2013; 28(Suppl. 1):8–12.
Please cite this article as: Feng C, et al, Transamniotic stem cell therapy (TRASCET) in a leporine model of gastroschisis, J Pediatr Surg (2016), http://dx.doi.org/10.1016/j.jpedsurg.2016.10.016
Contents lists available at ScienceDirect
Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg
Transamniotic stem cell therapy (TRASCET) in a leporine model of gastroschisis☆ Christina Feng, Christopher D Graham, Hester F. Shieh, Joseph A Brazzo III, John Patrick Connors, Lucas Rohrer, Alexander Papadakis, David Zurakowski, Dario O Fauza ⁎ Department of Surgery, Boston Children's Hospital and Harvard Medical School, Boston, MA
a r t i c l e
i n f o
Article history: Received 27 September 2016 Accepted 20 October 2016 Available online xxxx Key words: Amniotic mesenchymal stem cells Transamniotic stem cell therapy Gastroschisis Fetal stem cells
a b s t r a c t Background/purpose: Transamniotic stem cell therapy (TRASCET) with amniotic fluid mesenchymal stem cells (afMSCs) has been shown to mitigate bowel damage in a rodent model of gastroschisis. As a prerequisite to clinical translation, we sought to study TRASCET in a larger animal model. Methods: New Zealand rabbit fetuses (n = 64) with surgically created gastroschisis were divided into three groups. One group (untreated) had no further manipulations. Two groups received volume-matched intraamniotic injections of either saline or a suspension of afMSCs. Nonmanipulated fetuses served as controls. Histomorphologic measurements of intestinal damage, along with biochemical profiling of inflammation markers, were performed at term. Statistical comparisons were by Fisher's exact test, ANOVA and the Wald test (P b 0.05). Results: Overall survival was 62.5%. Segmental and total intestinal wall thicknesses were significantly decreased in the afMSC group compared with the untreated and saline groups (all P b 0.001), with no significant differences between untreated and saline groups (P = 0.24 to 1.00, depending on layer). Muscularis and serosal layers were significantly thicker in the afMSC group than in normal controls (P = 0.045 and P b 0.001, respectively). Conclusions: Concentrated intraamniotic injection of afMSC lessens, yet does not prevent, intestinal damage in a leporine model of gastroschisis. TRASCET may become a valuable strategy in the management of gastroschisis. Level of evidence: N/A - animal/experimental studies. © 2016 Elsevier Inc. All rights reserved.
Transamniotic stem cell therapy (TRASCET) has recently emerged experimentally as a practical paradigm for the treatment of different congenital anomalies [1–3]. It is based on the amplification of a biological role played by naturally occurring amniotic fluid mesenchymal stem cells (afMSCs), which we have previously shown to contribute to normal fetal tissue repair [4]. The administration of concentrated suspensions of afMSCs into the amniotic cavity, as per the TRASCET principle, has been shown to decrease intestinal damage in a rodent model of gastroschisis [3]. The eventual clinical translation of this therapeutic concept necessarily demands testing its efficacy in larger animal models. In this study, we sought to study TRASCET in a leporine model of gastroschisis.
☆ Authors' contributions:Data acquisition: CF, CDG, HS, JAB, JPC, LR, AP, and DOF.Analysis and data interpretation: CF, CDG, HS, DZ, and DOF.Drafting of the manuscript: CF, DZ, and DOF.Critical revision: CF, DZ, and DOF. ⁎ Corresponding author at: Boston Children's Hospital, Department of Surgery, 300 Longwood Avenue - Fegan 3, Boston, MA 02115. Tel.: +1 617 919 2966; fax: +1 617 730 0910. E-mail address: [email protected] (D.O. Fauza).
1. Methods This study was approved by the Boston Children's Hospital Institutional Animal Care and Use Committee under protocol #14-11-2852. 1.1. Donor afMSC processing Donor cells consisted of previously banked afMSCs obtained from normal heterogeneic New Zealand does time-dated at 23–24 days of gestation (E23–24; term = 32–33 days). Cells were isolated and expanded based on previously described methods [5]. Fluorescenceactivated cell sorting (FACS) analysis was used to confirm a mesenchymal progenitor identity using the Vantage SE cell sorter (BD Biosciences, San Jose, CA) and unconjugated mouse monoclonal antibodies previously validated for use in rabbits, namely CD44 (Antigenix America, Inc., Huntington Station, NY), CD45 (Antigenix), and CD90 (Life Technologies/Thermo Fisher, Carlsbad, CA). A mouse isotype immunoglobulin control was used to exclude nonspecific staining. Phenotyped and expanded cells were labeled by fluorescent nanocrystal technology using the Qtracker® cell labeling kit 525 (Life Technologies, Chicago, IL) emitting a wavelength of 525 nm with excitation at 405–485 nm, as per the manufacturer's instructions [6]. Green fluorescence was observed in up to approximately 90% of cells
http://dx.doi.org/10.1016/j.jpedsurg.2016.10.016 0022-3468/© 2016 Elsevier Inc. All rights reserved.
Please cite this article as: Feng C, et al, Transamniotic stem cell therapy (TRASCET) in a leporine model of gastroschisis, J Pediatr Surg (2016), http://dx.doi.org/10.1016/j.jpedsurg.2016.10.016
2
C. Feng et al. / Journal of Pediatric Surgery xxx (2016) xxx–xxx
postincubation using an EVOS® FL Color Imaging System microscope fitted with an on-board computer and integrated imaging software (Life Technologies, Carlsbad, CA). 1.2. Gastroschisis creation and treatment Eighteen time-dated pregnant New Zealand white does were housed individually under standard dark/light cycling conditions and fed a normal diet ad libitum. On E23, they were anesthetized with 1.5% to 2.5% isoflurane (Abbot Laboratories, North Chicago, IL) in 100% oxygen and given 20 mg/kg of cefazolin intravenously. Using sterile technique, the bicornuate uterus was partially exposed through a midline laparotomy. Depending on the number of fetuses present, one to two fetuses on the most distal aspect of each uterine horn were manipulated (i.e. 2–4 fetuses per doe). Under 2.5× magnification, a 1-cm hysterotomy was made, previously bordered by four equally spaced 5-0 Prolene (Ethicon, Somerville, NJ) stay sutures incorporating the myometrium and gestational membranes. Both hind limbs were exteriorized, while the remainder of the fetus was kept within the amniotic sac. A 3-mm full thickness abdominal defect was then created at either the right or the left lower quadrant using micro surgery scissors. This was followed by gentle blunt probing of the incision, so as to dilate the abdominal opening and allow for evisceration of bowel loops. Once evisceration was confirmed, the fetus was completely returned to the amniotic cavity and the hysterotomy was closed with running 5-0 Prolene suture (Ethicon). The uterus was returned to the abdomen and the laparotomy was closed in two layers with 3-0 Vicryl (Ethicon) and 5-0 Monocryl (Ethicon) running sutures. Animals were allowed to recover and proceed to term. A total of 64 fetuses with surgically created gastroschisis were divided into three groups. One group underwent no further manipulations (n = 22). Two treatment groups received volume-matched (2–3 mL) intraamniotic injections of either plain phosphate-buffered saline (PBS, n = 22) or a suspension of the previously processed heterologous rabbit afMSCs, at 2 × 106 cells/mL in PBS (n = 20), after a comparable amount of amniotic fluid had been removed from the amniotic cavity. The intraamniotic infusions were performed just prior to completion of the closure of the hysterotomy using an 18-gauge angiocatheter (Becton Dickinson, Franklin Lakes, NJ). The saline group was required for two reasons. One is the fact that the afMSCs were delivered in a saline suspension, which by itself could have had an effect. The other is the previous demonstration by other groups that saline alone may minimize bowel damage in the rabbit and other animal models, which has served as basis for amnioexchange having been attempted clinically. An additional group of nonmanipulated fetuses was used as normal controls (n = 10). 1.3. Histological analysis All does were euthanized just prior to delivery on E32–33 using 110 mg/kg of Fatal-Plus™ solution (Vortech, Dearborn, MI). The laparotomy was reopened and the uterus exposed. The amniotic cavities of the manipulated fetuses were incised and the fetuses inspected for the presence of gastroschisis. When present, the eviscerated portion of the intestine was excised en bloc. Approximately half of the specimen was fixed in 4% formaldehyde, paraffin embedded, and processed for regular histology with hematoxylin–eosin (H&E) staining. The other half was processed for biochemical analyses (details below). For nonmanipulated fetuses (normal group), a midline laparotomy was made and the intestines were removed en bloc from the pylorus to the rectum. Similarly in these fetuses, a portion of the normal intestine was fixed for H&E staining and another portion was processed for biochemical analyses. All histological analyses were performed using an EVOS® XL Core Imaging System microscope fitted with an onboard computer and integrated imaging software (Life Technologies). Quantitative histomorphologic measurements previously validated as surrogates of bowel damage in experimental gastroschisis [7,8] were
performed by three examiners blinded to the groups. Segmental thickness measurements of intestinal mucosa, muscularis, and serosa layers were performed separately in pixels at 200 × magnification using Image J software (NIH, Bethesda, MD) on at least four cardinal sections of axial slides of multiple intestinal loops. Four to 13 bowel loops per fetus were individually analyzed. Screening for labeled cell engraftment was performed in unstained slides using an EVOS® FL Color Imaging System microscope fitted with an on-board computer and integrated imaging software (Life Technologies). 1.4. Biochemical analysis Additional intestinal tissue samples from all groups were frozen at −80 °C at the time of procurement for measurements of myeloperoxidase (MPO), malondialdehyde (MDA), and interferon-gamma (IFN-γ) activities – all markers of inflammation. For the MPO and MDA colorimetry assays, samples were homogenized using the Bullet Blender Storm 24 (Next Advance, Averill Park, NY) and measured for absorbance at 412 nm and 532 nm, respectively, using a FLUOstar Omega microplate reader (BMG Labtech, Cary, NC) as per the manufacturer's protocol (Sigma-Aldrich, St Louis, MO). For the IFN-γ assay, sample homogenates were additionally prepared with 4 volumes of Dulbecco's PBS without Ca 2 + and Mg 2 + (Sigma-Aldrich) and 0.1% Triton™ X-100 (Fisher BioReagents, Pittsburgh, PA). Absorbance was at 450 nm on the FLUOstar Omega microplate reader using the RayBio® Rabbit IFN-γ ELISA kit per the manufacturer's protocol (RayBiotech, Inc., Norcross, GA). 1.5. Statistical analysis Fetal survival rates were compared using Fisher's exact test for binary proportions. Multiple bowel loops from each of the four comparison groups (normal, untreated, saline, and afMSC) were used in nested analysis of variance (ANOVA) comparisons of segmental and total intestinal wall layer thicknesses. A generalized estimating equation (GEE) approach with repeated measures was used in order to account for correlated data, e.g. multiple thickness measurements within the same fetus sample. Statistical significance was determined using the Wald test. Thickness data conformed to a normal Gaussian-shaped distribution, and thus, thickness measurements were presented using mean and standard error of the mean (SEM). Two-tailed values of P b 0.05 were considered statistically significant. Data analysis was performed using IBM SPSS software (SPSS version 23.0, IBM, Armonk, NY). 2. Results Of the 64 fetuses that underwent gastroschisis creation, 40 were viable at euthanasia (overall survival of 62.5%), of which 36 had a gastroschisis defect (Fig. 1). There was a statistically significantly higher rate of survival in the saline group (n = 20/22, 90%) when compared with the untreated (n = 11/22, 50%; P = 0.007) and the afMSC (n = 9/20, 45%; P = 0.002) groups. Survival was not significantly different between the untreated and afMSC groups (P = 0.77). Histomorphologic comparisons were performed in a total of 256 bowel loops distributed as follows: 87 loops in the normal control group, 67 in the untreated group, 54 in the saline group, and 48 in the afMSC group. There were no significant differences between the groups in the median number of bowel loops analyzed per fetus (P = 0.23). Segmental intestinal thicknesses of serosal, muscularis, and mucosal layers, as well as total intestinal wall thickness, were all significantly decreased in the afMSC group compared to those of the untreated (P b 0.001) and saline (P b 0.001) groups (Figs. 2 and 3). Total intestinal wall thickness and mucosal thickness of the afMSC group were not significantly different from those of normal controls (P = 0.26 and P = 1.00, respectively); however, both the muscularis and the serosal layers were increased in the afMSC group when compared to normal controls
Please cite this article as: Feng C, et al, Transamniotic stem cell therapy (TRASCET) in a leporine model of gastroschisis, J Pediatr Surg (2016), http://dx.doi.org/10.1016/j.jpedsurg.2016.10.016
C. Feng et al. / Journal of Pediatric Surgery xxx (2016) xxx–xxx
3
Fig. 1. Representative gross views of rabbit fetuses with gastroschisis at euthanasia. (A) Untreated group; (B) saline-treated group; and (C) afMSC-treated group.
(P = 0.045 and P b 0.001, respectively). Comparisons between the saline and untreated groups showed no significant differences in mucosa (P = 0.24), muscularis (P = 0.33), serosa (P = 0.90), and total (P = 1.00) intestinal wall thickness. Both the untreated and the saline groups had significantly increased segmental (all layers) and
total intestinal wall thickness compared to normal controls (P b 0.001 for all comparisons). Nanocrystal-labeled afMSCs were visualized sparsely throughout the intestine in the afMSC group, albeit not in the luminal epithelium (Fig. 3).
Fig. 2. Histological segmental and total intestinal wall thickness measurements were significantly decreased in the afMSC group compared with the untreated and saline groups (all P b 0.001), with no significant differences between untreated and saline groups (P = 0.24 to 1.00, depending on layer). However, muscularis and serosal layers were significantly thicker in the afMSC group than in normal controls (P = 0.045 and P b 0.001, respectively).
Please cite this article as: Feng C, et al, Transamniotic stem cell therapy (TRASCET) in a leporine model of gastroschisis, J Pediatr Surg (2016), http://dx.doi.org/10.1016/j.jpedsurg.2016.10.016
4
C. Feng et al. / Journal of Pediatric Surgery xxx (2016) xxx–xxx
Fig. 3. Representative histological views of the exposed intestine in (A) untreated, (B) saline, and (C) afMSC groups. The total bowel wall, serosal, muscular, and mucosal thicknesses appeared different in the afMSC group when compared with the untreated and saline groups, as reflected in the quantitative measurements shown in Fig. 2. However, they were not completely normal, as illustrated in (D) normal fetal intestine. H&E, 200× magnification. (E and F) Nanocrystal-labeled afMSCs were visualized sparsely throughout the intestine, except for the mucosal epithelium, in the cell-treated group. Unstained, 100× and 200× magnifications.
There were no significant differences in MPO (P = 0.82), MDA (P = 0.86), and IFN-γ (P = 0.17) activities across all groups. Specifically, concentrations detected were as follows, described as median values with interquartile range. For MPO - normal controls: 0 (0–19) mU/mL; untreated: 9 (0–16) mU/mL; saline: 3 (0–17) mU/mL; and afMSC: 13 (0–17) mU/mL. For MDA - normal controls: 0.39 (035–0.56) nmol/uL; untreated: 0.4 (0.29–0.48) nmol/uL; saline: 0.34 (0.24–0.51) nmol/uL; and afMSC: 0.39 (0.22–0.54) nmol/uL. For IFN-γ - normal controls: 0.48 (0.44–0.53) ng/mL; untreated: 0.54 (0.49–0.58) ng/mL; saline: 0.5 (0.47–0.55) ng/mL; and afMSC: 0.56 (0.49–0.63) ng/mL. 3. Discussion The prevalence of gastroschisis has increased significantly over the past two decades, for reasons as yet unknown. The most recent largescale data show its prevalence at almost 5 per 10,000 live births [9]. While multiple advances in perinatal care have led to exceedingly low mortality from isolated gastroschisis, little has changed in the significantly high morbidity rates associated with this disease, long known to derive largely from the prolonged exposure of the eviscerated intestine to amniotic fluid [10–12]. To date, no prenatal intervention, including deliberate preterm delivery, has been clinically proven to lessen intestinal damage, improve bowel function, or decrease overall morbidity. Although several approaches to minimize intestinal damage have been reported experimentally, a cell-based strategy had yet to be described until our recent study on TRASCET in a rodent model [3]. The TRASCET concept is centered on a primarily quantitative amplification of a naturally occurring biological process, somewhat analogous to the principles governing different forms of blood transfusions. Certainly, clinical translation of such a concept could not be properly justified without it being tested and refined in larger animal models. Although the sheep model is arguably the ultimate preclinical proving ground for this application, a rapid transition from the rodent to the exceedingly costly and resource-intensive ovine model would be less sensible than a more progressive one using an intermediate species, such as rabbits. Rabbits are members of the taxonomic order Lagomorpha, widely considered as a large animal model. Importantly in this case, the native amniotic fluid volume of rabbits is several-fold larger than that of rodents. More specifically, the volume of intraamniotic injections required for TRASCET in rabbits is 30 to 60 times larger than
that required for rats, depending on the disease process studied. The fetal rabbit model we used was essentially based on the first one described by Dr. Jens Rosenkrantz and colleagues in the 1970s, with few modifications [13]. While the leporine model is cost effective and carries a reasonable fetal survival rate, the leporine gestation is too short to allow for the use of autologous afMSCs, hence our use of banked heterologous cells. We saw no evidence of inflammation or any untoward reaction derived from the use of heterologous cells, as it was the case in the similar heterologous rodent model. At the same time, the exposed bowel in afMSCtreated animals was not entirely normal. We cannot rule out the possibility that autologous cells could have an impact on such results. The reasons for the statistically significantly higher survival rate in the saline group are not entirely clear, yet they may have derived from the fact that most animals in that group were the last ones to be manipulated, therefore after the steeper portion of the surgical learning curve had passed. The lack of statistically significant differences in the comparisons of segmental intestinal wall layer thicknesses between the saline and untreated groups is in accordance with both our results in the rodent model and the clinical experience with amnioexchange. Patients submitted to amnioexchange in the third trimester have not shown improvements in the degree of intestinal damage, time to full enteral feeds, or length of hospitalization [14]. In the present study, afMSC-based TRASCET did not completely reverse the bowel damage associated with gastroschisis, yet it did minimize it. The underlying mechanisms behind this outcome remain to be determined. Mesenchymal stem cells have long been known for possessing immunomodulatory and anti-inflammatory activities, which may well be germane to the effects we observed [15–17]. Nonetheless, we did not see significant differences between the groups in the activity of the inflammatory markers measured. Although molecules and signals other than those measured may be at play, the absence of at least certain inflammatory changes is actually in accordance with other studies, which also did not show variations in selective markers of inflammation in the setting of gastroschisis [14,18]. Besides further investigation into other components of the inflammatory response, such as quantitative and qualitative analyses of leukocyte activity and of connective tissue proliferation and/or fibrin deposition, among others, scrutiny into the status of native intestinal stem cells should be
Please cite this article as: Feng C, et al, Transamniotic stem cell therapy (TRASCET) in a leporine model of gastroschisis, J Pediatr Surg (2016), http://dx.doi.org/10.1016/j.jpedsurg.2016.10.016
C. Feng et al. / Journal of Pediatric Surgery xxx (2016) xxx–xxx
included, given that administered stem cells, including MSCs, are known to often exert a chaperone effect upon the local stem cell population, independently of inflammatory mediation. At the same time, we had to prioritize previously validated methods of comparison of bowel damage in models of gastroschisis, which, to our knowledge, are based on the methods used in the present study, this being the rationale for the methodology used. It was not our goal, nor was it possible, to understand all the mechanisms behind the protection afforded by the afMSC delivery, which we do intend to continue to investigate. Furthermore, as much as the mechanisms of afMSC-derived protection of the exposed bowel have yet to be clarified, the donor cell homing routes in that setting also remain to be fully elucidated and may not necessarily involve only direct seeding from the amniotic cavity. Of note, in the rabbit model, donor cells had to be administered at the time of creation of the defect, therefore before any bowel damage had actually taken place. This may have explained the sparse afMSC engraftment pattern observed. The timing of this model also does not allow for repeat cell injections, which could be envisioned clinically as means to enhance therapeutic benefit. At the same time, it could also be argued that, despite the unavoidable short time afforded by this model, there was still an effect, even if not a full reversal of the bowel damage. We expect the rodent model of gastroschisis to continue to be instrumental to further developments and to be pursued in parallel to larger animal models. The latter are not conducive to mechanistic investigations because of the notorious unavailability of supplies for pathway-specific analyses. Importantly, this and the other previously reported animal models of this disease have not faithfully and consistently reproduced the unique component of prenatal narrowing of the fascial defect present in the human condition, which plays a significant role in the mechanism of bowel damage besides the chemical insult derived from the exposure to amniotic fluid. It remains unclear whether TRASCET could have any impact on that component of the pathophysiology of this disease. Notwithstanding the inherent limitations of the leporine model of gastroschisis, it constitutes a helpful next step toward the eventual clinical translation of TRASCET. The present results showing that concentrated intraamniotic injection of afMSCs lessens intestinal damage in that model add further support to the notion that TRASCET may become a valuable strategy in the management of gastroschisis.
5
References [1] Dionigi B, Ahmed A, Brazzo III J, et al. Partial or complete coverage of experimental spina bifida by simple intra-amniotic injection of concentrated amniotic mesenchymal stem cells. J Pediatr Surg 2015;50(1):69–73. [2] Dionigi B, Brazzo III JA, Ahmed A, et al. Trans-amniotic stem cell therapy (TRASCET) minimizes Chiari-II malformation in experimental spina bifida. J Pediatr Surg 2015; 50(6):1037–41. [3] Feng C, Graham CD, Connors JP, et al. Transamniotic stem cell therapy (TRASCET) mitigates bowel damage in a model of gastroschisis. J Pediatr Surg 2016;51(1): 56–61. [4] Klein JD, Turner CG, Steigman SA, et al. Amniotic mesenchymal stem cells enhance normal fetal wound healing. Stem Cells Dev 2011;20(6):969–76. [5] Klein JD, Fauza DO. Amniotic and placental mesenchymal stem cell isolation and culture. Methods Mol Biol 2011;698:75–88. [6] Muller-Borer BJ, Collins MC, Gunst PR, et al. Quantum dot labeling of mesenchymal stem cells. J Nanobiotechnol 2007;5:9. http://dx.doi.org/10.1186/1477-3155-5-9. [7] Bittencourt DG, Barreto MW, Franca WM, et al. Impact of corticosteroid on intestinal injury in a gastroschisis rat model: morphometric analysis. J Pediatr Surg 2006; 41(3):547–53. [8] Hakguder G, Olguner M, Gurel D, et al. Induction of fetal diuresis with intraamniotic furosemide injection reduces intestinal damage in a rat model of gastroschisis. Eur J Pediatr Surg 2011;21(3):183–7. [9] Jones AM, Isenburg J, Salemi JL, et al. Increasing prevalence of gastroschisis – 14 states, 1995-2012. MMWR Morb Mortal Wkly Rep 2016;65(2):23–6. [10] Langer JC, Bell JG, Castillo RO, et al. Etiology of intestinal damage in gastroschisis, II. Timing and reversibility of histological changes, mucosal function, and contractility. J Pediatr Surg 1990;25(11):1122–6. [11] Langer JC, Longaker MT, Crombleholme TM, et al. Etiology of intestinal damage in gastroschisis. I: effects of amniotic fluid exposure and bowel constriction in a fetal lamb model. J Pediatr Surg 1989;24(10):992–7. [12] Srinathan SK, Langer JC, Blennerhassett MG, et al. Etiology of intestinal damage in gastroschisis. III: morphometric analysis of the smooth muscle and submucosa. J Pediatr Surg 1995;30(3):379–83. [13] Sherman NJ, Asch MJ, Isaacs Jr H, et al. Experimental gastroschisis in the fetal rabbit. J Pediatr Surg 1973;8(2):165–9. [14] Midrio P, Stefanutti G, Mussap M, et al. Amnioexchange for fetuses with gastroschisis: is it effective? J Pediatr Surg 2007;42(5):777–82. [15] Barry F, Murphy M. Mesenchymal stem cells in joint disease and repair. Nat Rev Rheumatol 2013;9(10):584–94. [16] Phinney DG, Prockop DJ. Concise review: mesenchymal stem/multipotent stromal cells: the state of transdifferentiation and modes of tissue repair – current views. Stem Cells 2007;25(11):2896–902. [17] Wang Y, Chen X, Cao W, et al. Plasticity of mesenchymal stem cells in immunomodulation: pathological and therapeutic implications. Nat Immunol 2014;15(11):1009–16. [18] Goncalves FL, Bittencourt DG, Velloso LA, et al. Corticosteroid effect upon intestinal and hepatic interleukin profile in a gastroschisis rat model. Acta Cir Bras 2013; 28(Suppl. 1):8–12.
Please cite this article as: Feng C, et al, Transamniotic stem cell therapy (TRASCET) in a leporine model of gastroschisis, J Pediatr Surg (2016), http://dx.doi.org/10.1016/j.jpedsurg.2016.10.016